3.82 Surveillance cameras are very much part of life in the twenty-first century, ever since the foundations of their use were laid in the latter decades of the twentieth century. Evidence of images from security cameras can be very helpful in identifying the perpetrators of crimes. Such evidence has been admitted in English courts, mainly in criminal cases.1 The widespread availability of video-recorded and tape-recorded evidence has opened up the possibility that such evidence may be augmented with more advanced techniques, and the enhancement of the sounds or images, together with the use of more advanced techniques such as aural identification and facial mapping, can help to identify the parties in a recording.
1 A list that is not exhaustive includes: McShane (1978) 66 Cr App R 97; R v Fowden and White [1982]
Crim LR 588 (CA); R v Grimer [1982] Crim LR 674, 126 SJ 641 (CA); R v Dodson (Patrick); R v Williams (Danny Fitzalbert Williams) [1984] 1 WLR 971, (1984) 79 Cr App R 220; Stockwell (Christopher James) (1993) 97 Cr App R 260; Clarke (Robert Lee) [1995] 2 Cr App R 425; Clare (Richard), Peach (Nicholas William) [1995] 2 Cr App R 333; R v Feltis (Jeremy) [1996] EWCA Crim 776; R v Hookway [1999] Crim LR 750; R v Briddick [2001] EWCA Crim 984; Loveridge (William) [2001] EWCA Crim 973, [2001] 2 Cr App R 29. In this instance, the accused were recorded by video in the court, an act which was prohibited by s 41 of the Criminal Justice Act 1925, and the recording was also held to have infringed the rights of the accused under art 8 of the Human Rights Act 1998 – however, neither infringement was held to have interfered with the right to a fair trial (E. Goldstein, ‘Photographic and videotape evidence in the criminal courts of England and Canada’ [1987] Crim LR 384).
3.83 Before such evidence is used, there should be a careful examination1 of the technology in question. A good example of this judicial scrutiny is that done by Steyn LJ in Clarke (Robert Lee),2 where his Lordship analysed the technique of facial mapping3 by video superimposition. The court carefully considered the reliability of the underlying scientific techniques, noting that the techniques themselves could be fit for debate, and their improper use by an expert in the particular case could in turn affect the probative value of such evidence. It was only after it was satisfied on these two grounds that the identification evidence from the application of the technique was admitted.
1 The careful examination may be done in a trial within a trial, also called a ‘voir dire’.
2 [1995] 2 Cr App R 425, 430F.
3 Michael C Bromby, ‘At face value?’ (2003) NLJ Expert Witness Supplement 301, 302–4; R v Jung [2006] NSWSC 658.
3.84 Issues regarding the reliability and application of these techniques are very much for expert evidence, depending on the nature and sophistication of each technique.
But some guidance may be sought that stem from the best practices for handling electronic evidence. For instance, for evidential techniques that involve manipulating and enhancing digital imagery, Gregory Joseph has noted that the following steps must be taken before enhanced digital imagery can usefully be used:1
1. The original image needs to be properly authenticated.
2. The original image must remain intact to enable the original to be compared with the enhanced version.
3. The original image should be preserved in such a way that its integrity cannot be impugned.
4. The process of enhancement should be fully documented.
5. The process of enhancement should be carried out in such a way that the process can be repeated by the other party.
6. The enhanced images should be preserved in such a way that prevents it from being manipulated and thereby preserves its integrity.
1 Gregory P Joseph, Modern Visual Evidence (Law Journal Press 2009) 4.
3.85 Important lessons were also spelt out regarding the use of voice recognition technologies and techniques for identification purposes in R v Flynn and St John.1 In this case, the prosecution sought to identify the two appellants as conspirators of a robbery through voice recognition techniques. Before the robbery, the police secretly fitted a listening and transmitting device to one of the vehicles it was assumed (correctly) that the conspirators would use for the robbery. Four police officers testified that they
recognized the appellants’ voices from the 60 minutes of covert recording made by the device. The trial judge ruled admissible the evidence of the police officers and the transcripts of the recording and placed the evidence before the jury. The appellants challenged the decision of the trial judge to admit the voice recognition evidence of the officers and the judge’s failure to give an appropriate direction to this evidence.
1 [2008] EWCA Crim 970, [2008] 2 Cr App R 20, [2008] Crim LR 799.
3.86 In giving judgment on appeal, Gage LJ noted that there are two categories of voice recognition evidence: expert evidence using either auditory analysis or acoustic/
spectrographic analysis, or lay listener evidence, where the lay listener as a witness is required to possess some special knowledge of the suspect that enables him to recognize the suspect’s voice. Such witnesses may be close relatives or friends, but they may also be persons who acquire such familiarity by the frequency of their contact with the suspect. Gage LJ also noted that suspect identification by voice recognition is more difficult than visual identification, that voice identification by experts using sophisticated auditory, acoustic and spectrographic and that sophisticated auditory techniques is likely to be more reliable than identification by a lay listener, and that the quality of identification by a lay listener is highly variable. In addition, research has shown that a confident recognition by a lay listener of a familiar voice may nevertheless be wrong, because while an expert is able to draw up an overall profile of the individual’s speech patterns, in combination with instrumental analysis and reference research, a lay listener’s response is fundamentally opaque because he cannot know and has no way of explaining which aspects of the speaker’s speech patterns he is responding to, and has no way of assessing the significance of the individually observed features relative to the overall speech profile. This makes it more difficult to challenge the accuracy of his evidence.
3.87 For all these reasons, the Court of Appeal allowed the appeal, holding that the police officers as lay listeners had a limited opportunity to acquire familiarity with the appellants’ voices, and that the quality of the covert recording was poor. In contrast, both experts, one representing the prosecution and the other representing the appellants, were unable to recognize their voices, further casting doubt on the officers’
voice recognition evidence.
3.88 While R v Flynn and St John did not close the door on voice recognition evidence, in a paper by Gary Edmond, Kristy Martire and Mehera San Roquem, the authors suggest the following minimal safeguards required before the prosecution can seek to admit voice recognition evidence from lay listeners:
1. The process must be properly recorded, and the amount of time spent in contact with the defendant will be very relevant to the issue of familiarity.
2. The date and time spent by the police officer compiling a transcript of a covert recording must be recorded. If the police officer annotates the transcript with his views as to which person is speaking, that must be noted.
3. A police officer attempting the voice recognition exercise must do so without the aid of a transcript that bears another officer’s annotations of whom he believes is speaking.
4. It is highly desirable that a voice recognition exercise should be carried out by someone other than an officer investigating the offence.1
1 [2008] EWCA Crim 970, [53]; also the paper by Gary Edmond, Kristy Martire and Mehara San Roque, ‘Unsound law: issues with (“expert”) voice comparison evidence’ (2011) 35 Melbourne University Law Review 52.
3.89 These safeguards are certainly in line with the issues raised by Gage LJ in R v Flynn and St John, and highlight the care with which both the parties and the courts must observe when seeking to admit computer-generated and computer-augmented evidence, in order to safeguard the evidential process.